get_platform_for_env() in build.sh parses `pio project config
--json-output` with Python's strict json.load(), which rejects literal
control characters inside string values. Depending on the installed
pio version, a build_flags value can embed one there, breaking
env-name lookup with "Invalid control character" before the actual
build even starts. strict=False accepts it, same as a lenient JSON
consumer would for a non-conforming producer.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
CMD_SET_RADIO_PARAMS and the on-device Settings radio editors
(Freq/SF/BW/CR/Preset) called radio_driver.setParams() directly instead
of applyRepeaterRadio(), so editing the companion's own radio params
while relaying on a dedicated profile silently retuned the live radio
off that profile. Both now route through applyRepeaterRadio(), the
single decision point already used at boot and by the on-device
repeater toggle.
Also: overhear-suppression matched queued packets by payload hash
alone, so it could cancel a queued direct-route retransmit that
happened to share a hash with an overheard flood packet — now scoped
to flood-queued entries only. repeaterProfileValid()'s frequency floor
(100 MHz) is raised to 150 MHz to match the radio's actual validated
range. RepeaterScreen's preset-name lookup now uses the same
epsilon-tolerant float comparison as Settings instead of exact ==,
avoiding inconsistent "Custom" labeling between the two screens.
Dedupe: the BW-options table, the value-column x-offset math, and the
default-profile-seeding block had each drifted into two or three
independent copies (RepeaterScreen vs Settings, MyMesh constructor vs
DataStore.cpp upgrade path); factored into shared helpers in
RadioPresets.h / NodePrefs.h.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Enabling the repeater now defaults to a dedicated radio profile rather
than relaying on the companion's current frequency: same-network
repeating isn't the MeshCore community norm. The default profile is
seeded with a frequency in the same legal band as the companion's own
network (433/868/915 MHz, via defaultRepeaterFreqForBand()) rather than
a flat firmware constant, so it can't land outside what's allowed for
the operator's region.
Also fixes a gap where a true first boot (no prefs file yet) never hit
the DataStore.cpp migration fallback and stayed on repeater_use_profile=0.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
sbrk(0)-high-water counted freed-and-recycled blocks as permanently
used, understating free heap. mallinfo().uordblks reflects bytes
actually outstanding. Also strip the trailing space from the decoded
error-flag string so right-alignment sits flush.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The name-bar live-power readout was gated on the raw tx_apc pref, so after an
upgrade with Auto pwr previously on, it kept showing a power value while the
repeater was active — looking like APC was running even though it's forced off
and power is pinned to the ceiling. Gate it on apcActive() instead (now public),
matching the "--" lock in Settings. The radio page's TX line still shows the
actual power, as before.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
A repeater wants full, consistent TX power for relay reach, and APC's feedback
(own ACKs / own flood echoes) never fires on forwarded traffic — so leaving APC
on would relay at whatever reduced power it last trimmed to, or sit inert on a
dedicated repeater network. Force it off whenever client_repeat is on, mirroring
the power-save lock:
- apcActive() = tx_apc && !client_repeat gates every APC control site
(sample/failure/track).
- applyApc() pins power to the tx_power_dbm ceiling; called at boot, on the
on-device repeater toggle, and on the app's CMD_SET_RADIO_PARAMS.
- Settings "Auto pwr" shows "--" and blocks the toggle while repeating; the
user's pref is preserved and restored when the repeater is switched off.
Also refreshes the FEATURES repeater entry (Hold-Enter reset, Errors row, schema
0x10, the radio-setting locks and status indicator).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Status bar: blinking » indicator while relaying (new ICON_REPEATER), added to
the 1 s blink cadence alongside the advert/trail markers.
- Power-save lock: a repeater must hear all traffic, so duty-cycle RX is forced
off whenever client_repeat is on. Effective state = rx_powersave && !repeat,
applied at boot, in applyPowerSave(), on the on-device toggle, and on the app's
CMD_SET_RADIO_PARAMS. Settings shows "--" and blocks the toggle while repeating;
the user's pref is preserved and restored when the repeater is switched off.
- Diagnostics: reset now uses the standard Hold-Enter one-item action menu (Back
dismisses) instead of an Enter/Cancel popup; new "Errors" row decodes the
Dispatcher ERR_EVENT_* flags (F/C/R, or OK).
- RepeaterScreen: drop the live stats block (they live on Diagnostics).
- Min-SNR load clamp tightened to the UI's -20..10 range.
Docs updated (settings_screen, tools_screen, FEATURES).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Move the repeater controls out of Settings > Radio (where the indented sub-item
labels collided with the value column on a 128px OLED) onto their own full-width
Tools > Repeater screen, holding the toggle, the five flood-politeness filters,
and live forwarding stats (Forwarded / Pool free / Queue) together. Toggle text
standardised to ON/OFF throughout.
Adds an optional dedicated radio profile (Network: Current/Custom): enabling the
repeater can switch the radio to a preset/manual freq/bw/sf/cr and disabling
restores the companion's params (revert-on-disable). A profile equal to the
companion = same network; a different one = drop onto a separate repeater
network. MyMesh::applyRepeaterRadio() is the single decision point (boot + every
toggle/edit), gated by repeaterProfileValid(); the profile re-applies after a
reboot left with the repeater on. Persisted behind schema 0xC0DE0010, cleared to
off when an upgraded file yields an invalid config.
Docs updated (settings_screen, tools_screen, README, FEATURES).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- settings_screen: full Radio section (Preset/Freq/SF/BW/CR/Repeater were
undocumented), plus the five flood-only politeness sub-items.
- tools_screen: new Diagnostics section (per-type counters, heap/stack, radio,
pool/queue, Forwarded, reset popup).
- README + FEATURES: mention diagnostics; FEATURES gets a full repeater entry.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When enabled (Settings > Radio > Suppress dup), a received flood packet whose
hash matches one still waiting in the outbound queue cancels that queued
retransmit: another node already relayed it, so this node stays quiet. Cuts
redundant airtime in dense meshes; pairs with Yield (longer delay = wider window
to overhear a peer first).
- Dispatcher: wantsOverhearSuppress() hook (default off) + suppressQueuedDuplicate()
scan over the outbound queue by packet hash; onRetransmitCancelled() hook.
- Mesh: overrides onRetransmitCancelled() to back out n_forwarded so the count
reflects what actually goes on air. resetStats() already covered.
- Flood-only and hash-matched (payload hash ignores path for non-TRACE, so our
copy and the peer's relayed copy hash equal). Own sends can't be overheard
before their first TX, so suppression only ever cancels forwards.
- New repeat_suppress_dup pref, schema sentinel 0xC0DE000F, stray-byte clamp.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Four opt-in filters, all default off, consulted only when the repeater is on
(MyMesh::allowPacketForward) and shown under Settings > Radio only while it is:
- Skip advert: don't re-flood ADVERT packets (highest-volume flood traffic).
- Max hops: drop a flood packet once it has travelled N hops.
- Yield: scale the FORWARDED flood retransmit delay so a mobile companion
defers to better-sited fixed repeaters; never touches the node's own sends.
- Min SNR: drop a flood copy received below a dB threshold so marginal fringe
traffic isn't re-flooded.
All four are flood-only — on a direct route this node is the named next hop, so
dropping there would kill delivery. Persisted behind a new schema sentinel
(0xC0DE000E) with stray-byte clamps; min-SNR uses a -128 "disabled" sentinel so
an upgraded file can't read as "filter at 0 dB".
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- DiagnosticsScreen: Enter opens a confirm popup (defaults to Cancel) that
zeroes the cumulative counters via Mesh::resetStats().
- Dispatcher::resetStats() made virtual; Mesh overrides it to also clear
n_forwarded so the reset covers the repeater forward count.
- REPEAT_LOOP_MAX gains a 4-byte-hash entry and isRepeatLooped() bounds-
checks hash_size, so it can't OOB-read if path_mode 3 is ever accepted.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
allowPacketForward() only gates permission; the packet can still be
dropped afterward (hasSeen(), path-size limit). Add Mesh::n_forwarded,
incremented at the 3 ACTION_RETRANSMIT_DELAYED return sites, so the
Diagnostics screen can confirm a repeater is actually relaying traffic
rather than just being configured to.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Shows packet counts by category (RX/TX), radio noise floor/RSSI/SNR,
packet-pool free count and outbound queue length, uptime, and heap/stack
headroom on a single scrollable screen.
Adds generic per-payload-type RX/TX counters to Dispatcher (mesh layer),
plus pool-free/queue-length getters, and a new DeviceDiag helper for
nRF52 heap (linker-symbol + sbrk) and stack (FreeRTOS high-water-mark)
stats — the only platform needed for the 3 in-scope boards (Wio Tracker
L1 OLED/Eink, GAT562 30S), all nRF52840.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New DigitEditor.h: standalone, reusable LEFT/RIGHT-cursor + UP/DOWN-spin
numeric editor. Replaces Freq's fixed ±0.025MHz nudge, which carried each
preset's own fractional offset forever, landing on a different unpredictable
grid depending which preset you started from. Every digit (hundreds down to
thousandths) is now directly addressable. SF/BW/CR are unchanged (small
discrete sets, plain left/right cycling is fine there).
- icons.h: drop the scroll-indicator's DARK contrast halo. Every caller
already keeps its selection bar clear of the indicator column (reserve/
_reserve), so the halo was never actually needed for contrast — and the
bottom arrow's halo had no clip symmetric to the top one, so it bled 1px
into the popup's bottom border.
- QuickMsgScreen.h: updateChannelUnread() now marks everything in the
rendered visible window as read, not just the highlighted row — a taller
screen that fits more message boxes at once should mark more read up
front. Also drops a premature call at CHANNEL_HIST entry that ran before
the channel's own render() had computed _hist_visible, which could ratchet
_viewing_max_seen past what was actually shown (that field is shared with
DM_HIST and stale until this channel's first render).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Matches every other list (drawList's row width - reserve): the highlight
shouldn't paint over the indicator's column, just frame it.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Tick marks, the scale baseline and the travel-direction pointer triangle
were hardcoded pixel constants, so they shrank to near-invisible slivers
next to the 2x cardinal labels and numeric readout on landscape e-ink.
Same fix pattern as the trail map / keyboard separator: derive sizes from
miniIconScale()/sepH() instead.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
PopupMenu drew its own unscaled ^/v text markers instead of the
track+thumb indicator the rest of the UI uses, so it didn't show position/
proportion and looked inconsistent on e-ink. drawScrollIndicator(Px) assumed
a full-width list (anchored to d.width()), which doesn't hold for a centred,
narrower popup box — added a right_x-aware overload so it can anchor to the
box's own edge instead.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Radio Preset popup: "+ Save current..." moved to the top of the list; add
"- Delete preset..." (shown only when a saved preset exists) which opens a
sub-list of just the saved presets and clears the picked slot.
- CMD_SET_RADIO_PARAMS no longer requires repeat=1 to land on one of the
three dedicated repeat_freq_ranges bands — none of the 16 built-in presets
matched those exactly, so app-driven repeat would either get rejected or
silently retune away from the user's chosen frequency. Commented out
(not deleted) so the old behaviour can be restored if needed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Settings > Radio gains:
- Repeater toggle — turns on flood forwarding (client_repeat) on the current
operating frequency, with loop-detection and an 8-hop advert cap ported
from simple_repeater's allowPacketForward.
- Packet pool bumped 16->32: forwarding's queued retransmits were starving
Dispatcher::checkRecv()'s allocNew(), silently dropping incoming DMs and
channel messages while repeating.
- Preset picker (16 community-suggested region presets) plus manual
Freq/SF/BW/CR fields for fine-tuning. Freq bounds come from the radio
driver's own RadioLib-validated range (SX1262: 150-960MHz) instead of a
guessed sanity bound.
- Save current radio settings as a named user preset (4 persisted slots),
alongside the built-in list.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
PopupMenu is compact and stands alone (no full-screen header to separate
from), so the gap meant for listStart() just wasted vertical space and let
the selection bar's edges miss the box border by a few px.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Trail map dots/diamonds/cross/start markers and the north indicator were
drawn as raw unscaled pixels, so they shrank to near-invisible specks on
landscape e-ink's larger font scale; they now route through the mini-icon
framework (icons.h) like the rest of the UI, and the grid intersection dots
scale with it too. The keyboard's preview/grid separator had the same bug
(hardcoded 1px) and now uses display.sepH().
Also makes UP/DOWN (and the keyboard's LEFT/RIGHT/UP/DOWN) wrap top<->bottom
across every list, menu and form selector, matching the behavior PopupMenu
already had.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Content drawn under the title separator touched the line on every standard
screen, because listStart() == headerH() (the row right after the separator).
Graphical screens worked around it with a hand-rolled hdr+2.
Bake a 2px breathing gap into listStart() so every list gains it at once, and
switch the screens that hand-rolled the offset (Compass, Nav, Nearby detail) to
listStart() so the content top is computed the same way everywhere.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The popup menu used fixed pixel constants (box width, item height, text
positions) sized for the OLED 128x64. On landscape e-ink the font is ~2x, so
rows overlapped, text ran past the box edge, and the box never adapted to its
contents.
Rebuild render() from live font metrics: box width fits the widest title/item
(clamped to the screen, centred), row height and title bar derive from the line
height, items and title are ellipsized so nothing overflows, and the scroll
arrows live in a reserved gutter. The title separator uses the scaled sepH()
with a 2px gap before the first row so content no longer touches the line.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The joystick rotation setting never survived a reboot: loadPrefsInt cleared it
on every load. The post-read override guarded on FEAT_JOYSTICK_ROTATION_SETTING
(and JOYSTICK_ROTATION), but DataStore.cpp included neither Features.h nor the
macro's header, so both were undefined — `#if !FEAT_JOYSTICK_ROTATION_SETTING`
was always true and the stored value was forced back to 0 every time.
Include Features.h so the flag resolves per build, and only force the default
when the setting isn't user-editable (OLED). On e-ink the stored value is now
kept; stale values migrated from another build are still corrected.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The solo build already copies firmware.zip (the adafruit-nrfutil DFU package
used for OTA / BLE-DFU updates) into out/, so it was present in the Actions
artifact but never in the tagged GitHub Release. Give it a friendly name
(solo-<ver>-<device>-ota.zip) and include out/*.zip in the release files.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Portrait e-ink message history had two scrollbar glitches:
- Track length tracked the last visible box's bottom, so a long message at
the bottom that left empty space shrank the whole bar. Pin the track to the
full list area (hist_start_y..cby); only the thumb sizes/moves.
- Thumb size was derived from the per-frame visible-box count, which fluctuates
with variable-height boxes. Drive it off pixel sums with a constant viewport
(view_px = track_h) so the thumb stays stable while scrolling one list.
- Gutter reserve used last frame's _hist_visible, so an incoming message briefly
toggled the gutter and reflowed box widths. Decide it from a whole-list fit
test at the widest layout instead — stable, no flicker.
Unify both fixes in computeHistScroll() (shared by DM + channel views) and
split drawScrollIndicator into a pixel core (drawScrollIndicatorPx) + an
item-count wrapper.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Fixes from a full ui-new audit (OLED + e-ink both build green):
- remove dead DisplayDriver::drawScrollArrows (all lists use drawScrollIndicator)
- SettingsScreen selection bar -> lineStep()-1, matching the drawList screens
- HomeScreen: don't force the 1s blink refresh on the CLOCK page (status icons
are hidden there anyway)
- FullscreenMsgView: clamp KEY_DOWN scroll to a cached _max_scroll instead of
over-incrementing and leaning on the next render to clamp
- DataStore: clamp use_lemon_font (>1 -> 0) on load, like the other enum fields
- move the ~1.5KB wrap scratch (trans/lines) off the render stack into shared
file-scope statics in FullscreenMsgView (single-threaded UI; the fullscreen
view and history list never lay out in the same frame)
Cleanup:
- drop dead members RingtoneEditorScreen::DUR_VALS and HomeScreen::sensors_scroll
- delete redundant manual scroll-clamps in handleInput across the drawList
screens (drawList already reclamps each render); remove the now-unused _visible
from QuickMsgScreen/NearbyScreen/BotScreen and BotScreen::scrollToSel()
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The DM/channel history views size the scroll-column reserve from the previous
frame's visible count. A new message bumps the count so the reserve briefly
appears until the next frame settles it — normally invisible, but the alert
overlay froze the screen (_next_refresh = _alert_expiry) for the alert's whole
duration, leaving the content shifted as if a scrollbar were needed. Keep the
underlying screen refreshing at its own cadence (capped at the alert expiry);
the display CRC skips unchanged frames so e-ink isn't thrashed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- drawList(): scrollable list skeleton (visible window, keep-sel-in-view,
scroll-column reserve, per-row callback, indicator). Migrate Tools, Bot,
Nearby, Waypoints list and the three QuickMsg lists onto it; centralises the
scroll-clamp/reserve/indicator that each had open-coded (and inconsistently).
- drawCenteredHeader(): centred title + separator, replacing the duplicated
two-liner across ~10 screens (incl. QuickMsg, which used width()/2 spacing and
was missed before); unifies the separator at headerH()-sepH().
- keyIsPrev()/keyIsNext() in UIScreen.h for the LEFT||PREV / RIGHT||NEXT idiom;
applied to 6 screens, behaviour-preserving. Cancel/context-menu stay
per-screen (they mean different things on different screens).
- rotLabel() dedups the 0/90/180/270 array in Settings.
- Unify selection-row height to lineStep()-1 (Tools/Bot matched the rest).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Extract drawBoxedIcon()/drawSlotIcon() in icons.h; collapse the four
near-identical mute/BT/advert/trail blocks in renderBatteryIndicator.
- Centre the glyph on the actual indicator box (box_h) instead of the text
line, fixing the 1px vertical offset in Lemon mode.
- Extract blinkOn() for the shared advert/trail blink cadence.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replace the old ^/v scroll arrows with the proportional scroll indicator
(thumb + triangle caps) used by the list screens. Reserve its right-edge
column and re-wrap the message text to the narrower width so nothing renders
under it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Replace single-letter M/B/A/G top-bar indicators with scalable mini-icons:
ICON_MUTE (speaker+cross), ICON_BLUETOOTH (rune), ICON_ADVERT (broadcast
mast+waves), ICON_TRAIL (map pin). Centred in the cw+2 indicator box;
disconnected BT shows the plain glyph instead of lowercase b.
- fix(settings): right-side values used display.valCol() without the scrollbar
reserve, so after expanding a section they rendered under the indicator.
Route all value cursors through valCol(display) = display.valCol() - _reserve.
- fix(settings): font toggle used ^=1, which on a stale value of 2 (older Hybrid
build) flips 2<->3 — both nonzero, locking applyFont() on Lemon with no way
back. Normalise: use_lemon_font = use_lemon_font ? 0 : 1.
- docs(settings): document Messages > Resend setting.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two independent, default-off toggles under Settings › Radio.
Pwr save: hardware RX duty-cycle (SX126x SetRxDutyCycle via
startReceiveDutyCycleAuto). The chip cycles RX↔sleep and wakes on a preamble —
no MCU state machine; recvRaw reads the packet exactly as in continuous RX.
Falls back to continuous RX on non-SX126x. (Replaces an earlier software-CAD
state machine that fought the hardware: polling a warm-sleeping chip gave a
phantom-busy channel that stalled TX ~4 s and dropped ACKs in the scan gaps.)
Auto pwr: Adaptive Power Control. tx_power_dbm becomes a ceiling; actual TX
power tracks the reverse-link SNR margin (measured above the per-SF demod floor,
EWMA-smoothed, proportional step with a deadband). Feedback comes from direct /
room-server ACKs and, for channels (no ACK), from hearing a repeater rebroadcast
our own flood; a lost confirmation ramps power back up so channel sends can't get
stranded below what the repeaters can hear.
Prefs schema 0xC0DE0009 (rx_powersave, tx_apc). Radio page / name bar show the
live TX power; noise floor reads n/a while duty-cycling.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Restore commented-out EXTERNAL_FLASH_DEVICES/USE_QSPI for reference;
board has no QSPI flash chip — defines were RAK4631 leftovers causing
framework to lock P0.28/P0.30 (joystick UP/LEFT) as QSPI outputs.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Board has no QSPI flash chip — EXTERNAL_FLASH_DEVICES/USE_QSPI were
copied from RAK4631. The framework was initialising the QSPI peripheral
on startup, locking P0.28/P0.30 (joystick UP/LEFT) as outputs. Removing
these defines prevents QSPI init entirely.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
QSPI IO0=P0.30 and IO2=P0.28 are wired to joystick LEFT and UP on this
board. QSPI init configures those pins as OUTPUT, blocking button reads.
Disable QSPIFLASH so the joystick works; firmware falls back to internal
flash storage.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Contacts with lastmod=0 or lastmod>now (RTC not synced after reboot but
contacts have timestamps from a prior session) were sorted to the top
because their raw lastmod > 0. Now both cases are normalised to 0 before
the comparison, so unknown-time entries always appear at the bottom.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- TIME mode: call refresh() every 3 s so newly-heard contacts bubble up
to the correct position; previously the list froze at the order captured
when the filter was selected.
- All filters: right-align the distance/age value within the right column
(width - textWidth - 2 px) instead of pinning it to a fixed left edge,
so short strings like "5s" or "3m" sit flush with the longer ones.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Trail.h gpxWaypoints/gpxPoint: snprintf truncation clamp was `> sizeof`
(writing the NUL terminator into the XML stream); corrected to
`>= sizeof - 1` at all 3 sites (<wpt>, <time>, <trkpt>).
- UITask.cpp battMvToPercent: raise 100% ceiling from 4200 → 4170 mV —
the board never charges to 4200, so the indicator was stuck at ~97%.
- Trail.h header comment: corrected RAM cost from "256×12 B=3 KB" to
actual CAPACITY(512)×16 B=8 KB.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
NearbyScreen:
- Hold Enter on a list row now shows Navigate and Save waypoint directly,
without entering the detail view first
- TIME added as the last filter slot (LEFT/RIGHT cycles Fav→ALL→…→Snsr→TIME);
time-sort shows age in right column and sorts by last-heard descending
- Save waypoint also available from detail view Options menu
UITask:
- Extract addWaypoint() helper (two overloads: with/without explicit timestamp)
consolidating the full()+add()+saveWaypoints()+alert pattern; TrailScreen
and QuickMsgScreen updated to use it
Fixes:
- SoundNotifier.h: correct NodePrefs.h include path (../NodePrefs.h)
- kbAddSensorPlaceholders: dereference _kb pointer in SettingsScreen,
QuickMsgScreen and BotScreen (signature changed to KeyboardWidget&)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Single-file Web Serial app for GitHub Pages. Connects to the device
over USB CDC, auto-detects <?xml…</gpx> frame from the Export GPX
action, and offers the file as a browser download with an editable
filename (defaults to trail_YYYY-MM-DD.gpx). Resets automatically
after download so back-to-back exports work without reconnecting.
Requires Chrome/Edge (Web Serial API). Enable GitHub Pages from the
repo settings pointing at the branch root or docs/ as needed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Deduplicate the snprintf+showAlert pattern used in three mark-as-read
sites into a single private method.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MODE_SELECT already showed the count; CONTACT_PICK ("Mark as read") and
CHANNEL_PICK ("Mark all read") were silently clearing unread without
any feedback. Count is read before zeroing so the alert is accurate.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Move all buzzer/melody dispatch (DM, channel, advert) into a new
SoundNotifier class in SoundNotifier.h, isolating fork-specific sound
code so upstream changes to notify() no longer cause merge conflicts.
The shared playSlot() helper also removes the custom_played
duplication that existed across the three cases.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The enum value and its case label were never emitted after PR #16 —
both dispatch sites in MyMesh.cpp now use advertReceivedFlood or
advertReceivedZeroHop exclusively.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Moves KeyboardWidget from a per-screen value member to a single instance
owned by UITask, passed to SettingsScreen, QuickMsgScreen, and BotScreen
as a pointer. Only one screen is ever active at a time, so the shared
state is safe. Saves ~1.5 KB of always-resident heap.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Reverts onDiscoveredContact to the upstream 4-param signature, adding a
separate virtual onDiscoveredAdvert(bool was_flood) with a default no-op
implementation. Only MyMesh overrides the new virtual; simple_secure_chat
and any future implementors need no changes on upstream merges.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
After PR #16 added None (slot=3) to melody options, only notif_melody_ad
had a bounds check. Add matching clamps for dm and ch to prevent stray
bytes from an older prefs file triggering buildMelodyFromPrefs with an
out-of-range slot.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Upstream's native NRF52 companion power-saving (PR #1238) and the preamble
16->32 bump for SF<9 both landed in the v1.16 merge. Update the battery/power
roadmap entry: the MCU now sleeps via board.sleep(0) when !hasPendingWork()
(no toggle — not sleeping only wastes power; it doesn't touch the radio), and
the preamble fix unblocks the CAD RX-windowing experiment on feat/power-saving.
Note the radio_set_*() globals were removed upstream (replace with radio_driver.*)
so that branch will need a rebase.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The splash shows the upstream base version from the hardcoded MESHCORE_VERSION
build flag (distinct from FIRMWARE_VERSION, which is our solo version). The
upstream merge bumped FIRMWARE_VERSION to v1.16-solo.0 but left this flag at
1.15. Update all three occurrences (wio + eink) to 1.16.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Self-review caught a regression I'd just introduced: feeding hasConnection()
from isClientConnected() (BLE *or* USB) broke the GPX-export collision warning
in TrailScreen, which relies on hasConnection() meaning "BLE app connected"
(BLE → USB dump is safe; otherwise the app may be on USB → warn).
Correct split:
- hasConnection() ← isBLEConnected() — BLE-specific consumers: BT status
indicator, pairing PIN, GPX-export warning.
- isClientConnected() (BLE bonded OR open USB-CDC port) used directly only by
the Auto buzzer mute and message-wake, which should react to either transport.
Builds clean on dual (OLED + e-ink) and pure-BLE.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
USB client presence IS detectable on nRF52 after all: (bool)Serial ==
tud_cdc_n_connected() (DTR — a host has the CDC port open). Add
isClientConnected() = BLE-bonded OR USB-CDC-open; DualSerialInterface
overrides it, base defaults to isConnected() (single-transport unchanged).
Wire it so each consumer gets the right signal:
- hasConnection() ← isClientConnected(): Auto buzzer mute + message-wake now
trigger on BLE or an open USB port (PR #14's intent, done correctly), but
not on charging-only (no host → DTR low).
- BT status indicator + pairing PIN stay on isBLEConnected() (BLE-specific).
Caveat: a plain serial monitor also asserts DTR, so it counts as connected.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
PR #14 (advert sound) changed setHasConnection() from isBLEConnected() to
isConnected() to also mute on USB — but DualSerialInterface::isConnected() is
hardcoded true (USB is always a send fallback) and ArduinoSerialInterface's is
too ("no way of knowing"), so USB-client presence is undetectable. The result
on every dual build: hasConnection() always true, which broke everything that
means "a companion app is connected":
- Auto buzzer mute → permanently muted (Auto behaved like Off)
- BT status indicator → always shown connected
- message-wake → display never woke for new messages
Revert to isBLEConnected() (the only detectable client signal). Mute-on-USB
isn't achievable; Buzzer = Off covers that case manually.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Bluetooth-page PIN prompt was gated on hasConnection(), which is fed by
the serial interface's isConnected(). DualSerialInterface::isConnected()
always returns true (USB always "connected"), so waiting_for_pair was always
false and the PIN was never drawn — on every dual build, i.e. the published
OLED + e-ink firmware. Pure-BLE builds were unaffected (there isConnected()
reflects bond state).
Gate the PIN on actual BLE-bonded state instead: add
AbstractUITask::isBLEConnected() (forwards to BaseSerialInterface::
isBLEConnected(), which DualSerialInterface overrides with the real BLE state
and pure-BLE inherits as its bond-aware isConnected()). The PIN now shows on
the Bluetooth page while BLE is on and not yet bonded.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Settings › Sound gains an "AD sound" row, and Auto-Advert notes the audible
"in range" heartbeat: the device chirps on every received advert (any node),
melody chosen by AD sound; there's no off, so silence it via the Buzzer
setting. built-in still chirps.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a navigation tagline to the intro and a dedicated "GPS navigation"
feature-highlights block (waypoints, compass, navigate-to-anything,
location sharing, trail, global units); note location nav in the Messages
entry and expand the Tools entry to cover waypoints/compass/navigate.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Release notes covering the navigation suite (waypoints, GPS compass,
navigate-to-point, mesh location share, waypoints in GPX), the global
metric/imperial Units setting, the long-message buffer fix, plus the
existing v1.14 items, under docs/releases/solo-v1.14.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- geo::parseLatLon now rounds lat/lon in double (matching the Add-by-coords
form) instead of through a float cast that lost ~1 m near 1e8.
- NearbyScreen: preserve the ping menu's selected row across a rebuild
instead of snapping back to "Send" when a result line arrives.
- NearbyScreen: guard the ping-menu input path so a vanished contact
(selectedStoredPubKey == false) closes the popup rather than acting on an
uninitialised pub-key buffer.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
When a trail exists but the framed box collapses to one point (e.g. a
single recorded point, or all points coincident, with no GPS fix), the
degenerate branch was drawing waypoint[0] at the centre — but waypoints
aren't folded into the box in trail-priority mode, so a possibly-distant
waypoint appeared at the trail's location. Only centre a waypoint in the
genuine "lone waypoint, no trail" case.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
GPX dumps previously contained only the <trk>. Emit saved waypoints as
<wpt> elements (label → <name>, plus <time> when the RTC was set) before
the track, per GPX 1.1 ordering, in both live and saved-file exports. The
emitter is duck-typed over the waypoint store so Trail.h stays decoupled
from Waypoint.h; labels are XML-escaped. They now import as pins in OsmAnd /
Garmin / GPX Studio alongside the track.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Capture the plan to multiply battery life via radio power technique (the
dominant draw is continuous RX, not the framework): CAD-based RX windowing
(~10-15mA → ~3-5mA, behind a setting due to the latency tradeoff) and
Adaptive Power Control, inspired by the ZephCore Zephyr port but stayable in
this Arduino tree. GPS power gating explicitly excluded — the device is a
live navigator/trail recorder. Deferred until trail polishing is done; prefer
adopting upstream nRF52 powersaving work.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The form's keyboard is the full alphanumeric one, and atof() silently read
a non-numeric entry as 0.0 — so typing letters created a waypoint at
(0, 0). Validate each coordinate with strtod and require the whole field to
be numeric; bad input now reports "Lat/Lon invalid" and saves nothing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Waypoints list header now reads "WAYPOINTS <n>/<cap>" (e.g. 3/16) so
it's clear at a glance how many waypoints are stored and how much room is
left before the table is full.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add-by-coords sat in the top-level Trail action popup alongside Mark here.
Move it into the Waypoints view as a trailing "+ Add by coords" row, so the
popup stays focused on trail/marking actions and waypoint management lives
in one place. "Waypoints" is now always shown in the action menu (it hosts
the list, backtrack and the add entry), so the first waypoint can be added
by coordinates even with no trail or saved points yet.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The on-screen keyboard has no comma or minus, so a single "lat,lon" entry
was unworkable. Replace it with a WP_ADD form: separate Lat / Lon / Label
rows plus Save. Lat/Lon take a magnitude from the keyboard and a hemisphere
(N/S, E/W) toggled with LEFT/RIGHT, so negative coordinates need no minus
key. Save validates the range, auto-names a blank label, and stores the
point via the normal waypoint path.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New Trail action "Add by coords" creates a waypoint without a GPS fix: the
keyboard takes a "lat,lon" string (decimal degrees), parsed by the shared
geo::parseLatLon; on success it falls through to the normal label keyboard
(blank → auto WP<n>). Useful for a meeting point or a spot read off a map.
Invalid input reports "Bad coordinates".
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Range test → ✅: shipped as the Nearby Nodes ping (RTT + SNR, repeatable);
the dedicated continuous-ping/sparkline screen was not built (kept as future).
- Lock-screen unread count → ✅: total "<n> unread" badge on the lock screen.
- Mark-read at type level → ✅ pointer to the implemented priority entry.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Waypoints/nav, GPS trail, Backtrack, Compass-to-contact flipped to ✅ with a
consolidated "shipped" summary noting deltas from the original spec
(two-char labels, heading tape compass, global Units, mesh location sharing,
trail-priority map framing, shared geo/gfx helpers).
- New backlog item: heading-up (track-up) map orientation, with the COG-only /
jitter / fit / grid / e-ink caveats and a lightweight rotate-the-fit plan.
- Audit: record the message-buffer length fix (history/compose sized to
MAX_TEXT_LEN) as ✅.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
While a trail exists the map now fits the recorded route (plus the live
position) and no longer folds waypoints into the bounding box, so a far-off
waypoint can't zoom the view out and squash the trail. Waypoints that
project outside the framed area are clamped to the nearest edge (marker +
label) instead of being lost off-screen. With no trail, the view still
auto-fits to the waypoints as before.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The label was always drawn up-and-right of the marker (wx+4, wy-3) and was
dropped entirely if it would clip the right edge — so waypoints near the
top/bottom (and left) edges had their label spill off the map or vanish.
Add an edge-aware placement helper: prefer upper-right, flip to the left
when it would clip the right edge, and clamp vertically so edge/corner
waypoints keep a visible label. Used by both the normal and degenerate
single-point map paths.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
After widening the history buffers, the compose side was still the
bottleneck: the keyboard capped input at 139 bytes and the quick-message
expand buffers were 140. Raise KB_MAX_LEN to 160 (MeshCore's MAX_TEXT_LEN)
and size the expand/compose buffers to MSG_TEXT_BUF, so a message can be
typed, expanded, sent, stored and displayed at the full protocol length on
every path.
The custom-message editor now passes an explicit max bound to its 140-byte
store, so the wider keyboard buffer can't overflow it; the bot editor
already passed per-field maxima.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Message history buffers were smaller than the over-the-air maximum, so long
messages lost their tail on display. Channel messages carry the sender
embedded as "Name: body" in the payload (up to MAX_TEXT_LEN = 160 B), but
ChHistEntry::text was only 140 B; DmHistEntry::text was just 80 B — and with
Polish text each accented char is two UTF-8 bytes, so the effective limit was
roughly half the visible characters.
Size both history buffers (and the fullscreen/preview split copies) to
MAX_TEXT_LEN + 1 so a full-length message is stored and shown intact. Sender
split buffers widened to the 32-char name limit.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Previously a shared location was only actionable after opening the message
fullscreen; Hold Enter on a history row only ever offered Reply. Route the
DM and channel list context menus through the same buildFsMenu /
dispatchFsAction path as the fullscreen view, so Navigate / Save waypoint
appear directly on the list row when the selected message carries a
location (Navigate returns to the list on Back).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Waypoints list → Hold Enter → Send hands the point to the Messages screen
as "[WAY]<lat>,<lon> <label>" (same text format geo::parseLatLon already
reads). The user picks a contact or channel, the text lands prefilled in
the keyboard to confirm/edit, then sends — closing the loop with the
Navigate / Save waypoint actions on the receiving end.
- QuickMsgScreen: share-compose mode (startShare/beginShareCompose). Picking
a recipient jumps straight to the prefilled keyboard; cancel returns home;
afterSend clears the mode.
- UITask::shareToMessage hands off from TrailScreen to the Messages screen.
- TrailScreen: "Send" added to the waypoint Rename/Delete popup; builds the
[WAY] payload at the {loc} precision (5 dp).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Locations already travel as plain text ({loc} expands to "lat,lon"), so any
message carrying coordinates can now be acted on. In the fullscreen message
view, the context-menu (Hold Enter) gains Navigate and Save waypoint
entries whenever the message contains a coordinate — alongside Reply.
- geo::parseLatLon: shared scanner for an embedded "lat,lon" (the {loc}
format), with an optional [WAY] tag that supplies a label. Range-checked,
requires a decimal point to avoid matching plain integer pairs. Will also
back the waypoint-share feature later.
- QuickMsgScreen: Navigate opens NavView inline (To/Hdg/distance, honouring
the global Units setting) over the message, back returns to it; Save
waypoint stores the point (label from [WAY], else auto-named). Works in
both DM and channel histories, and on incoming or outgoing messages.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The shrunken ring left only a few-pixel needle in the OLED's short vertical
space. Switch to a linear heading tape: a fixed travel-direction pointer at
centre over a scrolling N..E..S..W scale (30° ticks, N/E/S/W labelled),
with the large degrees+cardinal readout below. The tape is purely linear
(no trig) and uses the available width, so it stays readable on the small
panel. Update the Compass docs to match.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The N/E/S/W letters rode on the ring radius, so they overlapped the circle
as the card rotated. Place the letters on an outer radius and draw the ring
a glyph-height smaller, leaving a clear gap between them. The forward index
now sits in that gap, just outside the ring.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The no-heading hint was centred on the compass circle, overlapping the
ring. Draw it in the bottom readout band instead so the ring stays clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow the global Units preference in the two remaining Trail distance
spots that were still metric-only:
- Min dist gate: the index is now a unit-agnostic level (0=finest..3).
Metric stays 5/10/25/100 m; imperial uses round 15/30/75/300 ft (applied
as their metre equivalents in the sampling gate).
- Map scale grid: imperial mode uses round ft/mi steps (10 ft … 100 mi)
instead of metric steps with a converted label. Step storage moved to
float metres so sub-metre imperial steps keep their geometry; labels come
from a parallel table so the scale bar reads an exact round value.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add the Settings › System "Units" (Metric/Imperial) row and update the
Trail action-menu entry from the old four-way km/h·mph·min/km·min/mi
"Units" cycle to a Speed/Pace "Readout" toggle within the global system.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a single "Units" choice (Settings › System: Metric / Imperial) that
governs every distance and speed shown in Tools — Nearby Nodes, the Trail
views, and the navigate-to-point screen.
- NodePrefs: units_imperial (global) + trail_show_pace (Trail readout mode),
schema bumped 0xC0DE0005 → 0006. The old combined trail_units_idx is kept
for file-layout stability but no longer drives anything; upgraders default
to metric + speed.
- geo::fmtDist gains an imperial mode (ft below ~1000 ft, then miles).
- NavView / Nearby / Trail (waypoint list, Summary distance, List deltas,
map scale-bar label) all format through the global preference.
- Trail's action-menu "Units" row becomes a Speed/Pace readout toggle; the
km-vs-mi unit now follows the global setting instead of being one of four
combined options.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The last stray getLocationProvider/isValid duplicate (with its own
ENV_INCLUDE_GPS guard). Fold it into ownPos so every GPS lookup in the
nav screens goes through UITask::currentLocation.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The static haversineKm/bearingDeg/bearingCardinal/fmtDist members were
thin pass-throughs to GeoUtils. Remove them and qualify the call sites
with geo:: so there's no parallel local API to keep in sync.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
JakubandClaude Sonnet 4.6
